BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
First published online October 27, 2008
British Journal of Radiology (2009) 82, 49-61
© 2009 British Institute of Radiology
doi: 10.1259/bjr/58088207
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by OSEI, E K
Right arrow Articles by PANJWANI, D
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by OSEI, E K
Right arrow Articles by PANJWANI, D
British Journal of Radiology 82 (2009),49-61 ©2009 The British Institute of Radiology

Evaluation of daily online set-up errors and organ displacement uncertainty during conformal radiation treatment of the prostate

E K OSEI, BSc, MSc, PhD 1,2 R JIANG, PhD 1,2 R BARNETT, BSc, MSc, PhD 1,2 K FLEMING, BSc 3 and D PANJWANI, MD 3

1 Department of Medical Physics, Grand River Regional Cancer Center, 835 King Street West, Kitchener, 2 Department of Physics, University of Waterloo, 200 University Avenue West, Waterloo, 3 Radiation Treatment Program, Grand River Regional Cancer Center, 835 King Street West, Kitchener, Ontario, Canada

Correspondence: Ernest K Osei, Department of Medical Physics, Grand River Regional Cancer Center, 835 King Street West, Kitchener, Ontario, Canada. E-mail: ernestkwaku.osei{at}grhosp.on.ca


   Abstract
Top
 Abstract
Introduction
 Methods and materials
 Results
 Discussion
 Conclusions
 References
 
We have studied and analysed the magnitude of interfraction set-up errors and gold seed marker and prostate displacement in 118 patients using three gold seeds implanted within the prostate. Set-up errors and gold seed marker displacements were determined from bony anatomy and gold seed marker mismatch between the electronic portal image and the simulation digitally reconstructed radiograph (DRR), respectively. Prostate displacement relative to bony anatomy was determined from the difference between gold seed marker and bony anatomy displacement. Daily online repositioning of patients was accomplished through image matching using Varian Portal-Vision software. A total of 4878 electronic portal images and 236 DRRs from 118 patients were acquired over the course of the study. The means and standard deviations of the systematic error of gold seed marker displacement of 118 patients were 2.1±2.7 mm for anteroposterior (AP), –0.5±1.7 mm for left–right (L-R), and 1.0±1.9 mm for superoinferior (SI) directions; the random errors were 3.2 mm (0.9–4.9 mm) for AP, 1.9 mm (0.7–5.3 mm) for L-R, and 2.1 mm (0.7–4.5 mm) for SI directions. The mean and standard deviation of the isocentre set-up systematic error of 20 patients was 1.2±2.2 mm for AP, –0.1±1.4 mm for L-R, and –0.8±2.6 mm for SI directions. The isocentre set-up random errors were 1.6 mm (1.2–4.8 mm) for AP, 1.3 mm (0.6–2.5 mm) for L-R and 1.3 mm (1.0–2.6 mm) for SI directions. The mean and standard deviation of the prostate displacement systematic error relative to bony anatomy was 0.0±1.4 mm for AP, 0.0±1.1 mm for L-R and –0.2±2.4 mm for SI directions. Prostate displacement random errors were 1.5 mm (1.2–3.3 mm) for AP, 0.9 mm (0.4–1.5 mm) for L-R and 1.4 mm (1.2–2.4 mm) for SI directions.


   Introduction
Top
 Abstract
 Introduction
Methods and materials
 Results
 Discussion
 Conclusions
 References
 
Modern radiation therapy procedures increasingly utilize three-dimensional (3D) conformal delivery techniques, which necessitate more accurate patient positioning and tumour targeting. An improved dose delivery precision technique, such as intensity modulated radiation therapy (IMRT), necessitates a higher degree of accuracy in targeting the tumour (target delineation) and aligning the patient so as to reproducibly confirm that the tumour is in the same place in 3D space as the approved conformal dose distribution throughout therapy (positional verification). Therefore, the minimization of set-up errors and localization of the prostate within the patient have become critical components of conformal radiotherapy and IMRT of prostate cancer with escalated dose. Several approaches [1–40] have been taken to reduce the impact of position variation in prostate patients or to characterize the magnitude and nature of prostate movement. Yan et al [1] used an adaptive model to determine patient-specific variations and margins, and to remove systematic offsets in position. Ten Haken et al [4] have described prostate movement as a result of bladder and rectal filling. Kilovoltage or megavoltage cone beam CT (kV–CBCT or MV–CBCT) or kilovoltage on-board imaging (kV–OBI) systems are currently being employed for target volume localization in some cancer centres.

An overall increase in radiation dose delivered to the prostate tends to result in a greater overall survival rate [11]; however, as the prostate gland lies in close proximity to the rectal wall and bladder, any volume increase necessary to account for the combination of patient set-up error, as well as prostate movement, increases the volume of the rectal wall and bladder exposed to high doses of radiation and may pose an increased risk of radiation-induced injury [12]. Therefore, to gain full advantage of the dosimetric benefit of conformal and IMRT techniques, continual efforts are made to reduce target-positioning errors throughout the course of radiation therapy treatment [13]. Daily alignment with lasers on external skin marks or tattoos and a weekly set of portal films for bony anatomy alignment have become part of a standard technique to reduce set-up errors during radiotherapy. However, the outer patient skin marks are not an accurate fiducial reference for internal structures, and hence more accurate position verification is based on internal fiducials. The use of radio-opaque gold seed markers implanted in the prostate and visualized daily using electronic portal image devices and, more recently, kV-OBI (or in some centres with kV-CBCT) has therefore been employed to quantify organ position variation and provide an accurate and efficient method for prostate localization [14–16]. Seed marker migration has been studied by several investigators to quantify and characterize the movements of the seeds within the prostate; it has been established that the seeds do not migrate significantly within the prostate [14, 15].

The development of conformal radiation therapy and the expected therapeutic gain, as well as the increased risk of exposing critical organs near the target to very high doses, has always necessitated accurate definition of the margins around the clinical target volume (CTV). In order to accurately determine the margin required around the CTV based on our centre's practices, we instituted a study to investigate both isocentre set-up errors and prostate motions. The purpose of this paper is therefore to report on our study and analysis of the magnitude of interfraction isocentre set-up errors, and prostate and seed marker displacement, in patients who were treated with image-guided radiotherapy at our centre using three gold seeds implanted within the prostate. We compared the use of skin markers for standard patient position verification with image-based verification on internal structures, such as the bony anatomy and seeds implanted into the prostate.


   Methods and materials
Top
 Abstract
 Introduction
 Methods and materials
Results
 Discussion
 Conclusions
 References
 
Data from 118 patients were used to study gold seed marker displacements relative to the isocentre, and data from 20 patients were used to study bony anatomy displacement relative to the isocentre and prostate displacement relative to bony landmarks. We define isocentre set-up error or seed marker displacement as the difference between the measured position of a bony anatomy or seed marker (as defined by an electronic portal image (EPI)) and the "original" position (as defined by a digitally reconstructed radiograph (DRR)), respectively. Prostate displacement is the difference between the seed marker position relative to the isocentre and the bony anatomy position relative to the isocentre for the same patient. To quantify isocentre set-up errors and seed marker and prostate displacements, statistical data such as the mean, standard deviation (SD) and minimum and maximum displacements are calculated. Comparing the bony anatomy and seed markers on both the DRR and portal images requires that both images are represented in the same coordinate system, and we use the projection of the isocentre of the treatment beam chosen during treatment planning as the origin.

The mean isocentre set-up error and gold seed marker and prostate displacement for a patient was calculated by averaging over all fractions (Equation 1), where n is the number of fractions for each patient, dj is the daily isocenter set-up error (b), or seed marker (s) or prostate (p) displacement, and ms,b,p is the mean displacement for each patient:


Formula 001

The associated standard deviation ({sigma}(s,b,p)) is also calculated. The mean displacement and isocentre set-up error of the entire group was then computed by averaging over all patients (Equation 2), where p is the number of patients and Ms,b,p is the mean displacement of the entire group:


Formula 002

The root mean square (RMS) of the standard deviations from all patients is used to determine the standard deviation ({sigma}RMS) of the entire group (Equation 3):


Formula 003

We distinguish between random and systematic errors. For each patient, the mean value of the daily isocentre set-up error and seed marker or prostate displacement is the patient's systematic isocentre set-up error and seed marker or prostate displacement, respectively. The ensemble of differences between the isocentre set-up error and seed marker or prostate displacements and their means (i.e. standard deviations) comprises the distribution of random errors for the isocentre set-up error and seed marker or prostate displacements for the patient.

Patient preparation
Our centre initiated a Phase I/II prospective study of pure hypofractionated radiation treatment of prostate cancer. Low-risk patients (T1B/T1C/T2A and Gleason score ≤6, and prostate-specific antigen (PSA) ≤10) receive 5000 cGy in 15 fractions over 7 weeks (treatment delivered twice a week); intermediate-risk patients (T2B or Gleason score 7, or PSA >10) receive 6000 cGy in 20 fractions over 8 weeks (treatment delivered twice a week, alternating with thrice a week). The goal of the study was to reduce acute and late side-effects without compromising local tumour control. Dose escalation was proven to be effective for this disease and, consequently, image-guided radiation therapy plays an integral role in preparing these patients for accurate reproducible treatments with daily online image verification. Patients with intermediate- and low-risk adenocarcinoma of the prostate were identified for radiation treatment in this clinical trial. The eligibility criteria included clinical stage T1B, T1C and T2, N0, M0, a Gleason score <8 and PSA levels ≤20. The Tri-Hospital Research Ethics Board approved the trial and all patients signed a letter of information and consent to participate in the trial. To enable the visualization of the prostate during daily imaging, these patients had three gold seeds implanted transrectally in the prostate via a biopsy needle under ultrasound guidance. The striated soft-tissue gold seed markers were cylindrical in shape, and measured 1.2 mm in diameter by 3 mm in length (Northwest Medical Physics Equipment; Orange City, IA). Transrectal implantation of the gold seed markers in the prostate is similar to performing a prostate biopsy, with the possible side effects being rectal bleeding and infection [17].

CT planning
To allow the swelling from the seed implantation to subside, the planning CT scan (Philips AcQsim CT; Philips Medical Systems, Cleveland, OH) was acquired 1 week after implantation, and treatment usually began a few days later. Patients were instructed to arrive for the CT procedure with a full bladder (by drinking two glasses of water approximately 1 h before the appointment and not voiding) and with an empty rectum (asked to void before the appointment). Patients lay in a supine position on a solid flat carbon-fibre couch top and a sturdy foam immobilization device was placed below the knees. The pelvis area was scanned at 3 mm intervals and with 3 mm slice thicknesses, from the level of the fifth lumbar vertebra to the anal orifice. Three tattoos were then placed on the anterior, right and left lateral pelvis after the CT scan.

Radiation treatment planning
All patients' treatment planning was performed using the Pinnacle Treatment Planning System (Philips Pinnacle version 7.4 TPS; Milpitas, CA). When the CT images had been uploaded onto the treatment planning system, the prostate gland (i.e. the CTV) was contoured by the radiation oncologist. The planning target volume (PTV) was drawn by expanding the CTV by 7 mm on the posterior aspect and 10 mm in all other directions. The organs at risk, which, included the rectum (from the level of the inferior ischial tuberosity to the rectosigmoid flexure), bladder and femoral heads, were also contoured. A 3D conformal four-field box treatment plan was generated using 15 MV photon beams. The acceptable minimum and maximum doses to the PTV were 95% and 105%, respectively, of the prescribed dose. Dose constraints were employed for the rectum, bladder and femoral heads. The dose to 25% of the rectum should not be in excess of 48.5 Gy in 15 fractions or 52.5 Gy in 20 fractions. After the treatment planning, one pair of orthogonal (anteroposterior (AP) and right-lateral) DRRs was constructed in which the position of the seed markers and bony anatomy could be located. This image set constituted the reference image pair and were exported from the treatment planning system to the treatment station, together with the patient treatment plan.

Patient repositioning before each treatment fraction
The patients were again instructed to arrive with a full bladder (by drinking two glasses of water approximately 1 h prior to the appointment and not voiding) and with an empty rectum (asked to void before each appointment) to replicate the conditions for obtaining the DRRs. The patients were immobilized and treated in the supine position. On each treatment day, patients were positioned using laser alignment to the marks on their skin. The therapist acquired one pair of orthogonal electronic portal images using the AP and left-right (L-R) set-up fields; these constituted the comparison image set. The portal images are taken using a Varian Oncology Systems a-Si flat panel electronic portal imager (Varian Medical Systems, Palo Alto, CA) mounted on a dual energy Clinac 2100EX accelerator. The imager has a sensitive area of 512 x 384 pixels, with a pixel size of 0.784 mm. All images were acquired using three monitor units. Using both pairs of images, the daily repositioning of the patients before each treatment fraction (calculation of the repositioning displacements) was accomplished through image matching using Varian Portal-Vision softwareTM installed on the Varian's four-dimensional treatment console computer. This software aligns the three selected seed-marker positions on the EPI with the same positions on the DRR, and then determines the required patient repositioning displacements. A total of 4878 EPIs and 236 DRRs from 118 patients were acquired over the course of the study. These data were used to analyse seed marker displacement relative to the isocentre.

Organ displacement determination for 20 patients
In order to study the displacement of bony anatomy relative to the isocentre, and the displacement of the prostate relative to bony anatomy, we used in-house localization software (Proloc) to study the data of the first 20 patients, comprising 40 simulated DRRs and 1284 EPIs. The software is capable of determining displacement using either seeds markers or bony anatomy. In "seed mode", the three positions of the projected seed markers were selected on both the DRR and the EPI; the software then aligns the selected positions on the EPI to the same positions on the DRR, and determines the required displacement in the AP, L-R and superoinferior (SI) directions. Using the same set of patient images and in "bony anatomy mode", two reproducible bony anatomy points (usually the anterior pubic symphysis and the inferior pubic ramus or sacrum) were selected on both the DRR and EPI. The software once again aligns the points on the two sets of images and then determines the required displacement in the AP, L-R and SI directions.


   Results
Top
 Abstract
 Introduction
 Methods and materials
 Results
Discussion
 Conclusions
 References
 
gold seed marker displacement analysis: 118 patients
We assume that gold seed marker displacement encompasses both isocentre set-up error and prostate displacement. For a population of 118 patients, Table 1Go shows the mean systematic and random errors for seed marker displacements relative to the isocentre in the AP, L-R and SI directions. The mean systematic errors were 2.1±2.7 mm for AP, –0.5±1.7 mm for L-R and 1.0±1.9 mm for SI directions; the random errors were 3.2 mm for AP, 1.9 mm for L-R and 2.1 mm for SI directions. The data show a dominance of seed displacement in the posterior direction. The frequency distribution of the measured seed displacement shows a normal distribution in the L-R and SI directions. The distribution of seed marker displacement relative to the isocentre over the entire study (2439 fractions) is shown in Figure 1a–cGo. Figure 1aGo gives the distribution in the AP direction, Figure 1bGo in the L-R direction and Figure 1cGo in the SI directions. The AP distribution has more distributed points above zero (posterior direction), indicating a dominance in the posterior direction; the L-R and SI distribution has equally distributed points above and below zero, indicating no preferential direction of the displacements.


View this table:
[in this window]
[in a new window]

  		Table 1. Mean and standard deviation for systematic({Sigma}) and random ({sigma}) errors for seed-marker displacement relative to the isocentre in a population of 118 patients in the anteroposterior (AP), left-right (L-R) and superoinferior (SI) directions. The seed-marker systematic error for a patient was calculated by averaging all the seed marker displacements over all fractions, and the mean systematic error of the entire group was then computed by averaging over all patients. The root mean square (RMS) of the standard deviation from all patients is used to determine the random error ({sigma}) of the entire group

 

Figure 1
View larger version (27K):
[in this window]
[in a new window]

  		Figure 1. Distribution of gold seed marker displacement relative to the isocentre over the entire study for all 118 patients.(a) Anteroposterior: anterior ("–"), posterior ("+") direction. (b) Left-right: left ("–"), right ("+") direction. (c) Superoinferior: superior ("–"), inferior ("+") direction.

 
The cumulative distribution of seed marker displacement relative to the isocentre in the anterior, posterior, left, right, superior and inferior directions is shown in Figure 2Go. The cumulative distribution of the seed displacement was again largest in the posterior direction. Most of the displacements in the AP direction were found to be in the posterior direction (74%), compared with 26% in the anterior direction. In the AP direction, 49% of the displacements were ≤3 mm, 38% were >3 mm to ≤7 mm; 10% were >7 mm to ≤10 mm, and 3% were >10 mm. In the L-R direction, 83.6% of the displacements were ≤3 mm, 15% were >3 mm to ≤7 mm, 1% were >7 mm to ≤10 mm and 0.4% >10 mm. In the SI direction, 70% of the displacements were ≤3 mm, 25% were >3 mm to ≤7 mm, 4% were >7 mm to ≤10 mm and 1% were >10 mm. The individual patient systematic error, as a function of random error for the seed marker displacement relative to isocentre, is shown in Figure 3a–cGo. Figure 3a, 3b and 3cGo show the scatter plots in the anterior ("–")/posterior ("+") direction (Pearson correlation coefficient r = 0.22), the left ("–")/right ("+") direction (Pearson correlation coefficient r = 0.05) and the superior ("–")/inferior ("+") direction (Pearson correlation coefficient r = 0.05), respectively. The small correlation coefficients show that there is no significant correlation between the systematic and random errors of seed marker displacement.


Figure 2
View larger version (17K):
[in this window]
[in a new window]

  		Figure 2. Cumulative frequency distribution of seed marker displacement relative to the isocentre for 118 patients in the anterior, posterior, left, right, superior and inferior directions.

 

Figure 3
View larger version (15K):
[in this window]
[in a new window]

  		Figure 3. Scatter plot of the individual patient systematic error as a function of the random error for the seed marker displacement relative to the isocentre for 118 patients. The seed marker systematic error for a patient was calculated by averaging all of the seed marker displacements over all fractions, and the standard deviations are the random errors.(a) Anterior ("–")/posterior ("+") direction, Pearson correlation coefficient, r = –0.22. (b) Left ("–")/right ("+") direction, Pearson correlation coefficient, r = 0.05. (c) Superior ("–")/inferior ("+") direction, Pearson correlation coefficient, r = 0.05.

 
Prostate displacement and isocentre set-up error analysis: 20 patients
Statistical data (for 20 selected patients) comprising the mean systematic and random errors for (i) seed marker displacement (ii) isocentre set-up error and (iii) prostate displacement relative to bony anatomy is shown in Table 2Go.


View this table:
[in this window]
[in a new window]

  		Table 2. Mean and standard deviation for systematic({Sigma}) and random ({sigma}) errors for seed marker displacements relative to the isocentre, set-up error and prostate displacement relative to bony anatomy in a population of 20 patients in the anteroposterior (AP), left-right (L-R) and superoinferior (SI) directions. The systematic error for a patient was calculated by averaging all the displacements over all fractions, and the mean systematic error of the entire group was then computed by averaging over all patients. The root mean square (RMS) of the standard deviations from all patients is used to determine the random error ({sigma}) of the entire group

 
Isocentre set-up error
The cumulative frequency distribution of set-up errors of the patients is shown in Figure 4Go. The cumulative distribution of set-up error was least in the L-R direction. In the AP direction, 68% of the set-up errors were ≤3 mm, 28% were >3 mm to ≤7 mm, and 4% were >7 mm to ≤10 mm. In the L-R direction, 88.3% of the set-up errors were ≤3 mm, 11.2 % were <3 mm to ≤7 mm, and 0.5 % were >7 mm to ≤10 mm. In the SI direction, 78.1% of the displacements were ≤3 mm, 18.8 % were >3 mm to ≤7 mm, and 3.1% were >7 mm to ≤10 mm. No set-up error in any of the directions was >10 mm. The mean set-up errors were 1.2±2.2 mm for AP, –0.1±1.4 mm for L-R, and –0.8±2.6 mm for SI directions; the random errors were 1.6 mm for AP, 1.3 mm for L-R and 1.3 mm for SI directions (Table 2Go). When separated into the different directions, the mean set-up errors were 2.5±2.2 for anterior, 3.0±2.2 for posterior, –2.5±2.0 for superior, 2.2±1.6 for inferior, 1.7±1.3 for left and 1.8±1.4 for right directions. Frequency histograms of all set-up errors showed normal distributions for all three directions.


Figure 4
View larger version (19K):
[in this window]
[in a new window]

  		Figure 4. Cumulative frequency distribution of the set-up error for 20 selected patients in the anterior, posterior, left, right, superior, and inferior directions. The isocentre set-up error is the bony anatomy mismatch between the simulated digitally reconstructed radiograph and the daily electronic portal image.

 
Prostate displacement
The cumulative frequency distribution of prostate displacement is shown in Figure 5Go. The cumulative distribution of prostate displacement was least in the L-R direction. In the A-P direction, 83.5% of the prostate displacements were ≤3 mm, 15.3% were 3 mm to ≤7 mm, and 1.2% were >7 mm <to ≤10 mm. In the L-R direction, 96% of the set-up errors were ≤3 mm, and 4% were >3 mm to ≤7 mm. In the SI direction, 75.9% of the displacements were ≤3 mm, 20.6% were >3 mm to ≤7 mm, 3.1% were >7 mm to ≤10 mm and 0.4% were >10 mm. No prostate displacement in AP and LR directions was >10 mm. The mean systematic errors for prostate displacement were 0.0±1.4 mm for AP, 0.0±1.1 mm for L-R, and –0.2±2.4 mm for SI directions; the random errors were 1.5 mm for AP, 0.9 mm for L-R, and 1.4 mm for SI directions (Table 2Go). Frequency histograms of all prostate displacements showed normal distributions for all three directions. There is a larger set-up error than organ displacement in the AP direction, whereas a larger prostate motion than set-up error was found in the SI direction.


Figure 5
View larger version (20K):
[in this window]
[in a new window]

  		Figure 5. Cumulative frequency distribution of prostate displacement relative to the pelvis for the 20 selected patients in the anterior, posterior, left, right, superior, and inferior directions.

 
Isocentre set-up error vs seed marker displacement
Scatter plots of bony anatomy vs seed marker displacement of the 20 selected patients are shown in Figure 6a–eGo. All displacements were taken relative to the isocentre. The x-axis depicts displacements of the seed markers relative to the isocentre, whereas the y-axis shows the set-up error relative to the isocentre. Figures 6a, 6b and 6cGo show the scatter of the displacements in the AP, L-R and SI directions, respectively. Figure 6dGo shows the individual patient systematic errors and Figure 6eGo shows the random displacements in the AP, L-R and SI directions. The correlation between seed marker displacement and set-up error is highest in the L-R direction (Pearson's correlation coefficient, r = 0.83) and least in the SI direction (r = 0.58). In the AP direction, the correlation coefficient is 0.78. The correlation coefficients between systematic displacement of bony landmark and seed marker are 0.87 in the AP direction, 0.76 in the L-R direction and 0.56 in the SI direction A similar correlation of the random error gave correlation coefficients of 0.83, 0.92 and 0.55 in AP, L-R and SI directions, respectively.


Figure 6
View larger version (15K):
[in this window]
[in a new window]

  		Figure 6. Scatter plots of bony landmarkvs seed marker displacement. All displacements are relative to the isocentre. (a) Anterior ("–")/posterior ("+") direction, Pearson correlation coefficient, r = 0.78. (b) Left ("–")/right ("+") direction, Pearson correlation coefficient, r = 0.83. (c) Superior ("–")/inferior ("+") direction, Pearson correlation coefficient, r = 0.59. (d) The systematic displacement in the anteroposterior (AP), left-right (L-R) and superoinferior (SI) directions; r = 0.87, 0.76 and 0.60 in AP, L-R and SI directions, respectively. (e) The random errors in the AP, L-R and SI directions; r = 0.83, 0.92 and 0.55 in AP, LR and SI directions, respectively.

 
Prostate displacement vs isocentre set-up error
The individual patient systematic prostate displacement and set-up errors are shown in Figure 7a–cGo. The error bars indicate one standard deviation (1 SD). The AP, L-R and SI systematic prostate displacements for each patient are calculated by averaging over all fractions. The systematic error of prostate displacement varied from –3.2 mm to 2.4 mm for AP, –1.7 mm to 3.4 mm for L-R, and –4.4 mm to 7.2 mm for SI directions; the random errors varied from 1.2 mm to 3.3 mm for AP, 0.4 mm to 1.5 mm for L-R, and 1.2 mm to 2.4 mm for SI directions. The systematic set-up errors varied from –3.6 mm to 4.4 mm for AP, –2.2 mm to 2.3 mm for L-R, and –7.6 mm to 3.7 mm for SI directions; the random errors varied from 1.2 mm to 4.8 mm for AP, 0.6 mm to 2.5 mm for L-R, and 1.0 mm to 2.6 mm for SI directions. The data indicated that there is no correlation between set-up error and prostate displacement. The set-up errors in the AP direction were frequently larger in magnitude than those in the SI and L-R directions, and smallest in the L-R direction. The dominant set-up error in the AP direction was in the posterior direction, and in the SI was in the superior direction. A comparison of cumulative frequency distribution of set-up error and prostate displacements in the anterior, posterior, inferior, superior, left and right directions is shown in Figure 8a–cGo. There are larger set-up errors than organ displacements in the posterior and superior directions, whereas larger prostate motion than set-up errors was found in the anterior and inferior directions.


Figure 7
View larger version (16K):
[in this window]
[in a new window]

  		Figure 7. The systematic isocentre set-up error and prostate displacement for the 20 selected patients in the anteroposterior, left-right and superoinferior directions. The error bars indicate one standard deviation (1 SD). (a) Anterior ("–")/posterior ("+") direction. (b) Left ("–")/right ("+") direction. (c) Superior ("–")/inferior ("+") direction.

 

Figure 8
View larger version (17K):
[in this window]
[in a new window]

  		Figure 8. A comparison of the cumulative frequency distribution for set-up errors and prostate displacement in 20 selected patients. (a) Anteroposterior directions; (b) left and right directions; and (c) superoinferior directions.

 
A scatter plot of prostate displacement and set-up errors is shown in Figure 9a–eGo. It shows the relationship between internal prostate displacement relative to the pelvis, and bony anatomy displacement relative to the isocentre. For bony anatomy displacements, posterior displacement overweighs the anterior displacement, and no significant preferences were found in the SI and L-R directions. For prostate displacement, no preferential displacement was observed from the scattered distribution, but the lateral prostate displacement was smallest among the three directions. In the SI direction, the scattered points in the second quadrant show some larger bony anatomy displacements in the superior direction, whereas the prostate moves in the inferior direction relative to the pelvis. There is no correlation between the prostate motion and the pelvic motion.


Figure 9
View larger version (17K):
[in this window]
[in a new window]

  		Figure 9. Scatter plots of prostate displacement relative to pelvisvs bone anatomy displacement relative to the isocentre. (a) Anterior ("–")/posterior ("+") direction, Pearson correlation coefficient, r = –0.03. (b) Left ("–")/right ("+") direction, Pearson correlation coefficient, r = –0.16. (c) Superior ("–")/inferior ("+") direction, Pearson correlation coefficient, r = –0.27. (d) The systematic displacement in the anteroposterior (AP), left-right (L-R) and superoinferior (SI) directions; r = 0.15, –0.12 and –0.34 in the AP, L-R and SI directions, respectively. (e) The random error in the AP, LR and SI directions; r = 0.38, 0.35 and 0.31 in the AP, L-R and SI directions, respectively.

 
The time-course trends of prostate displacement and set-up error for three arbitrarily selected patients in all three directions are illustrated in Figure 10a–cGo. These plots are representative of the distributions observed for all patients in the study, and linear regression analysis of individual patients shows no statistically significant trends in displacement as the treatment progressed.


Figure 10
View larger version (24K):
[in this window]
[in a new window]

  		Figure 10. Prostate displacement(corrected for set-up errors) and set-up errors over the entire course of treatment for three randomly selected patients in the anteroposterior (AP), left-right (L-R) and superoinferior (SI) directions.

 

   Discussion
Top
 Abstract
 Introduction
 Methods and materials
 Results
 Discussion
Conclusions
 References
 
Our data on gold seed marker displacement show a dominance in the posterior direction; this observation may be caused by the different couch tops used in the CT and treatment rooms. A uniform carbon-fibre couch top is used in the CT room, whereas a tennis-racket-type carbon-fibre couch top is used in the treatment room. We therefore suspect that patients who are treated on the tennis-racket couch top experience slight "anatomical sag" relative to the lateral tattoos, and could be part of the explanation for this observation. However, because of the uncertainties in the measurements, it would not be easy to prove that this is actually occurring. We also suspect that a reduction in the size (diameter) of the rectum (as patients experience the onset of diarrhoea) as radiation treatment progresses is also a contributory factor. Thus, as treatment progresses, the prostate (gold seed markers) may fall posterior owing to smaller rectal size. Again, this study would not be able to prove that this is actually the case. However, Pinkawa et al [41], who studied prostate position variability and dose–volume histograms in radiotherapy for prostate cancer with full and empty bladders, demonstrated that the prostate has a tendency for displacement towards the posterior direction and that this was due to a decreasing rectal volume during radiotherapy.

Our data for the seed-marker displacement, set-up errors and prostate displacement can be compared with others in the literature (GoTables 3 and 4Go). Some studies [22, 27, 28 , 32, 34, 37, 39] characterized set-up errors and seed marker and prostate displacements of the entire patient group by reporting the standard deviation (1 SD). Others [19, 20, 24, 33, 35, 36, 38, 40] reported the individual patient mean displacement and the standard deviations, the mean of those mean values and the spread ({Sigma}), and the RMS, or averaged quadratically the individual standard deviations from all patients to obtain the standard deviation ({sigma}) of the entire group. Table 3Go shows the comparison of our data with other authors who characterized the errors by 1 SD ({sigma}); Table 4Go shows a similar comparison but the errors are reported as the mean and the spread of the mean ({Sigma}).


View this table:
[in this window]
[in a new window]

  		Table 3. Comparison of seed marker, set-up error (bony anatomy) and prostate displacement by different authors who characterized errors by one standard deviation (1 SD)

 

View this table:
[in this window]
[in a new window]

  		Table 4. Comparison of set-up error (bony anatomy) and prostate displacement by different authors who characterized errors by the mean and the spread around the mean

 
Hanley et al [19] performed a retrospective analysis of port films from 50 patients. Patient positioning uncertainty for a given set-up was determined using port films from three projections: two oblique and one lateral. A total of 1239 port films and 300 simulator films were analysed for the study. The distribution of systematic set-up errors for the 50 patients had a mean and standard deviation of –0.1±1.9 mm, 0.4±1.4 mm and –0.3±1.3 mm in the left-right, SI and AP directions, respectively. The distribution of random set-up errors about the mean approximated a normal distribution, and the standard deviations for the population of patients in the L-R, SI and AP directions were 2.0 mm, 1.7 mm and 1.9 mm, respectively. Nederveen et al [33] analysed 2025 portal images from 23 patients, and reported the mean systematic prostate displacement to be 0.0±1.0 mm in the L-R direction, 0.0±2.3 mm in the AP direction and 1.0±4.1 mm in the SI direction. The mean systematic set-up error was 0.0±2.1 mm for L-R, –1.0±4.4 mm for AP, and 0.1±2.1 mm for SI directions. We obtained a correlation coefficient between seed marker displacement and set-up error of 0.83 in L-R direction, 0.58 in the SI direction and 0.78 in the AP direction; Nederveen et al [33] reported values of 0.71 in the AP direction, 0.92 in the L-R direction and 0.46 in the SI direction. We also obtained a correlation coefficient between the systematic displacement of bony landmark and seed marker of 0.87 in the AP direction, 0.76 in the L-R direction and 0.56 in the SI direction; Nederveen et al [33] reported values of 0.86 in the AP, 0.91 in the L-R, and 0.08 in the SI directions.

Alasti et al [40] also analysed a total of 2549 portal images from 33 patients. Data from 23 patients were analysed for set-up errors and 10 were analysed for prostate motion. Set-up errors were characterized by standard deviations of 1.8 mm for AP and 1.4 mm for SI directions, and displacements due to prostate motion of 5.8 mm for AP and 3.3 mm for SI directions. Little et al [31] studied 35 patients being treated for prostate cancer with IMRT. They underwent daily B-mode acquisition and targeting (BAT) ultrasound localization and weekly orthogonal portal imaging. They reviewed a total of 243 pairs of orthogonal portal films and the corresponding daily BAT images. The mean set-up shift ± SD in the L-R, AP and SI directions was 0.0±2.8 mm, –0.2±3.0 mm and –0.0±2.0 mm, respectively, for portal films, and –0.8±3.2 mm, –1.4±6.4 mm and –1.7±6.4 mm, respectively, for BAT images taken on the same day as the portal films. The mean prostate organ motion measurements were –0.89±3.3 mm (R-L), –1.3±5.7 mm (AP), and –1.6±6.4 mm (SI).

During treatment, patients were aligned on a flat couch using their skin marks or tattoos. The reproducibility of "true isocentre" is based on the agreement between the relative position of the anatomy at treatment and that of CT simulation. Ideally, all parts of the anatomy during treatment should be conformed to the CT simulation anatomy. This means that bony anatomy and soft tissue would be in the same position relative to the tattoos, and the prostate displacement relative to the pelvis should be zero for all of the fractions. However, prostate displacement occurs independently from the bony anatomy, and results in both systematic and random deviations. Comparison of the systematic and random errors shows that systematic errors are significantly larger in the AP and SI directions. In this work, set-up errors are found to be predominant in the AP direction, which may occur as a result of the use of skin marks to determine the isocentre height, in combination with the use of the pelvic bones as a match structure. This agrees with the work by Hanley et al [19], Nederveen et al [33], Schiffner et al [36] and Alasti et al [40], who all showed the largest displacement to be in the AP direction and the smallest in the L-R direction; however, the work is contrary to that by Altholf et al [22] and Rudat et al [39], whose work showed the largest displacement to be in the SI direction. The movement of the skin marks used for patient positioning relative to the pelvic bones may result in a set-up error. Skin movement might be caused by respiration, weight loss or relaxation of the patient. This movement is expected to be small in the SI and L-R directions, and more pronounced in the AP direction. The prostate is not attached directly to bony anatomy, and our data show that the prostate displacement differs from pelvic bony anatomic position. This indicates that repositioning the patient using bony anatomy only slightly contributes to better target positioning. However, the data also show that, in individual cases (see Figure 10Go), the deviation of the target position can increase by applying position verification that is based on bony anatomy. This agrees with previous work [6, 24, 33, 40] on the correlation between both target motion and set-up errors, which showed that the target motion is dominant over the set-up error. Most studies [6, 19–22, 24, 27, 28, 33, 34, 37, 40] investigating prostate organ motion are consistent with our study in reporting the largest displacement in the AP direction and the smallest movement in the L-R direction.


   Conclusions
Top
 Abstract
 Introduction
 Methods and materials
 Results
 Discussion
 Conclusions
References
 
Daily EPIs combined with gold seed markers continue to provide an objective method with which to verify and correct the position of targets immediately before radiation delivery. Its routine clinical use will continue to improve the precision of external beam radiation therapy. The systematic error for individual patients can increase after using bony anatomy-based position verification owing to prostate movement that is independent of bony anatomy, and hence margins are still needed to account for organ motion. Significant reductions in both systematic and random errors in prostate localization radiation therapy can be achieved with online target-based position verification using gold seed image-based verification or (recently) cone beam CT. Through this study, a protocol has been established for performing daily online set-up and target displacement analysis and correction using bony anatomy (for patients without gold seeds) and seed markers, respectively.

Received for publication October 17, 2007. Revision received February 19, 2008. Accepted for publication March 19, 2008.


   References
Top
 Abstract
 Introduction
 Methods and materials
 Results
 Discussion
 Conclusions
 References
 

  1. Yan D, Vicini F, Wong J, Martinez A. Adaptive radiation therapy. Phys Med Bio. 1997;42:123–32.
  2. Slater JD, Yonemoto LT, Rossi Jr CJ, Reyes-Molyneux NJ, Bush DA, Antoine JE, et al Conformal proton therapy for prostate carcinoma. Int J Radiat Oncol Biol Phys 1998;42:299–304.[Medline]
  3. Lattanzi J, McNeeley S, Pinover W, Horwitz E, Das I, Schultheiss TE, et al. A comparison of daily CT localization to a daily ultrasound-based system in prostate cancer. Int J Radiat Oncol Biol Phys 1999;43:719–25.[Medline]
  4. Ten Haken RK, Forman JD, Heimburger DK, Gerhardsson A, McShan DL, Perez-Tamayo C, et al. Treatment planning issues related to prostate movement in response to differential filling of the rectum and bladder. Int J Radiat Oncol Biol Phys 1991;20:1317–24.[Medline]
  5. Balter JM, Sandler HM, Lam K, Bree RL, Lichter AS, Ten Haken RK. Measurement of prostate movement over the course of routine radiotherapy using implanted markers. Int J Radiat Oncol Biol Phys 1995;31:113–8.[Medline]
  6. Vigneault E, Pouliot J, Laverdiere J, Roy J, Dorion M. Electronic portal imaging device detection of radioopaque markers for the evaluation or prostate position during megavoltage irradiation: a clinical study. Int J Radiat Oncol Biol Phys 1997;37:205–12.[CrossRef][Medline]
  7. van Herk M, Bruce A, Kroes APG, Shouman T, Touw A, Lebesque JV. Quantification of organ motion during conformal radiotherapy of the prostate by three-dimensional image registration. Int. J Radiat Oncol Biol Phys 1995;33:1311–20.[CrossRef][Medline]
  8. Gall KP, Verhey LJ, Wagner M. Computer-assisted positioning of radiotherapy patients using implanted radiopaque fiducial. Med Phy 1993;20:1153–9.
  9. Zellars RC, Roberson PL, Strawderman M, Zhang D, Sandler HM, Ten Haken RK, et al. Prostate position late in the course of external beam therapy: patterns and predictors. Int J Radiat Oncol Biol Phys 2000;47:655–60.[Medline]
  10. Nederveen AJ, Dehnad H, van der Heide UA, van Moorselaar RJA, Hofman P, Lagendijk JJW. Comparison of megavoltage position verification for prostate irradiation based on bony anatomy and implanted fiducials. Radiother Oncol 2003;68:81–8.[Medline]
  11. Litzenberg D, Dawson LA, Sandler H, Sanda MG, McShan DL, Ten Haken RK, et al. Daily prostate targeting using implanted radiopaque markers. Int J Radiat Oncol Biol Phys 2002;52:699–703.[CrossRef][Medline]
  12. Zelefsky MJ, Leibel SA, Gaudin PB, Kutcher GJ, Fleshner NE, Venkatramen ES, et al. Dose escalation with three-dimensional conformal radiation therapy affects the outcome in prostate cancer. Int J Radiat Oncol Biol Phys 1998;41:491–500.[CrossRef][Medline]
  13. Storey MR, Pollack A, Zagars G, Smith L, Antolak J, Rosen I. Complications from radiotherapy dose escalation in prostate cancer: preliminary results of a randomized trial. Int J Radiat Oncol Biol Phys 2000;48:635–42.[CrossRef][Medline]
  14. Fraass BA, McShan DL, Kessler ML, Matrone GM, Lewis JD, Weaver TA. A computer-controlled conformal radiotherapy system I: Overview. Int J Radiat Oncol Biol Phys 1995;33:1139–57.[Medline]
  15. Sandler HM, Bree RL, McLaughlin PW, Grossman HB, Lichter AS. Localization of the prostatic apex for radiation therapy using implanted markers. Int J Radiat Oncol Biol Phys 1993;27:915–9.[Medline]
  16. Lam KL, Ten Haken RK. Improvement of precision in spatial localization of radio-opaque markers using the two-film technique. Med Phys 1991;18:1126–31.[Medline]
  17. Aubin M, Pouliot J, Milender L, Shinohara K, Pickett B, Anezinos C, et al. Daily prostate targeting with implanted gold markers and an a-Si flat panel EPID at UCSF: a five year clinical experience. Int J Radiat Oncol Biol Phys 2004;60 (Suppl):S266–7.
  18. Kutcher GJ, Mageras GS, Leibel SA. Control, correction, and modeling of setup errors and organ motion. Semin Radiat Oncol 1995;5:134–45.[CrossRef][Medline]
  19. Hanley J, Lumley MA, Mageras GS, Sun J, Zelefsky MJ, Leibel SA, et al . Measurement of patient positioning errors in three-dimensional conformal radiotherapy of the prostate. Int J Radiat Oncol Biol Phys 1997;37:435–44.[Medline]
  20. Roeske JC, Forman JD, Mesina CF, He T, Pelizarri CA, Fontenla E, et al. Evaluation of changes in the size and location of the prostate, seminal vesicles, bladder, and rectum during a course of external beam radiotherapy. Int J Radiat Oncol Biol Phys 1995;33:1321–9.[CrossRef][Medline]
  21. Beard CJ, Kijewski P, Bussiere M, Gelman R, Gladstone D, Schaffer K, et al. Analysis of prostate and seminal vesicle motion: Implication for treatment planning. Int J Radiat Oncol Biol Phys 1996;34:451–8.[CrossRef][Medline]
  22. Althof VGM, Hoekstra CJM, te Loo H-J. Variation in prostate position relative to adjacent bony anatomy. Int J Radiat Oncol Biol Phys 1996;34:709–15.[Medline]
  23. Langen KM, Jones DTL. Organ motion and its management. Int J Radiat Oncol Biol Phys 2001;50:265–78.[CrossRef][Medline]
  24. Wu J, Haycocks T, Alasti H, Ottewell G, Middlemiss N, Abdolell M, et al, Positioning errors and prostate motion during conformal prostate radiotherapy using on-line isocentre set-up verification and implanted prostate markers. Radiother Oncol 2001;61:127–33.[CrossRef][Medline]
  25. Van den Heuvel F, Fugazzi J, Seppi E, Forman JD. Clinical application of a repositioning scheme, using gold markers and electronic portal imaging. Radiother Oncol 2006;79:94–100.[Medline]
  26. Stroom JC, Heijmen BJM. Geometrical uncertainties, radiotherapy planning margins, and the ICRU-62 report. Radiother Oncol 2002;64:75–83.[Medline]
  27. Balter JM, Lam KL, Sandler HM, Littles JF, Bree RL, Ten Haken RK. Automated localization of the prostate at the time of treatment using implanted radiopaque markers; technical feasibility. Int J Radiat Oncol Biol Phys 1995;33:1281–6.[CrossRef][Medline]
  28. Litzenberg DW, Balter JM, Lam KL, Sandler HM, Ten Haken RK. Retrospective analysis of prostate cancer patients with implanted gold markers using off-line and adaptive therapy protocols. Int J Radiat Oncol Biol Phys 2005;63:123–33.[Medline]
  29. Scarbrough TJ, Golden NM, Ting JY, Fuller CD, Wong A, Kupelian PA, et al. Comparison of ultrasound and implanted seed marker prostate localization methods: Implications for image-guided radiotherapy. Int J Radiat Oncol Biol Phys 2006;65:378–87.[Medline]
  30. Chen J, Lee RJ, Handrahan D, Sause WT. Intensity-modulated radiotherapy using implanted fiducial markers with daily portal imaging: assessment of prostate organ motion. Int J Radiat Oncol Biol Phys 2007;68:912–9.[Medline]
  31. Little DJ, Dong L, Levy LB, Chandra A, Kuban DA. Use of portal images and BAT ultrasonography to measure setup error and organ motion for prostate IMRT: implications for treatment margins. Int J Radiat Oncol Biol Phys 2003;56:1218–24.[Medline]
  32. Antolak JA, Rosen II, Childress CH, Zagars GK, Pollack A. Prostate target volume variations during a course of radiotherapy. Int J Radiat Oncol Biol Phys 1998;42:661–72.[Medline]
  33. Nederveen AJ, van der Heide UA, Dehnad H, van Moorselaar RJA, Hofman P, Lagendijk JJW. Measurements and clinical consequences of prostate motion during a radiotherapy fraction. Int J Radiat Oncol Biol Phys 2002;53:206–14.[Medline]
  34. Melian E, Mageras GS, Fuks Z, Leibel SA, Niehaus A, Helen L, et al. Variation in prostate position: quantitation and implications for three-dimensional conformal treatment planning. Int J Radiat Oncol Biol Phys 1997;38:73–81.[CrossRef][Medline]
  35. Crook JM, Raymond Y, Salhani D, Yang H, Esche B. Prostate motion during standard radiotherapy as assessed by fiducial markers. Radiother Oncol 1995;37:35–42.[CrossRef][Medline]
  36. Schiffner DC, Gottschalk AR, Lometti M, Aubin M, Pouliot J, Speight J, et al. Daily electronic portal imaging of implanted gold seed fiducials in patients undergoing radiotherapy after radical prostatectomy. Int J Radiat Oncol Biol Phys 2007;67:610–9.[Medline]
  37. Van Herk M. Errors and margins in radiotherapy. Semin Radiat Oncol 2004;14:52–64.[CrossRef][Medline]
  38. Tinger A, Michalski JM, Cheng A, Low DA, Zhu R, Bosch WR, et al. A critical evaluation of the planning target volume for 3-D conformal radiotherapy of prostate cancer. Int J Radiat Oncol Biol Phys 1998;42:213–21.[Medline]
  39. Rudat V, Schraube P, Oetzel D, Zierhut D, Flentje M, Wannenmacher M. Combined error of patient positioning variability and prostate motion uncertainty in 3D conformal radiotherapy of localized prostate cancer. Int J Radiat Oncol Biol Phys 1996;35:1027–34.[Medline]
  40. Alasti H, Petric PM, Catton CN, Warde PR. Portal imaging for evaluation of daily on-line setup errors and off-line organ motion during conformal irradiation of carcinoma of the prostate. Int J Radiat Oncol Biol Phys 2001;49:869–84.[CrossRef][Medline]
  41. Pinkawa M, Asadpour B, Gagel B, Piroth MD, Holy R, Eble MJ. Prostate position variability and dose-volume histograms in radiotherapy for prostate cancer with full and empty bladder. Int J Radiat Oncol Biol Phys 2006;64:856–61.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by OSEI, E K
Right arrow Articles by PANJWANI, D
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by OSEI, E K
Right arrow Articles by PANJWANI, D

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
BJR DMFR IMAGING  ALL BIR JOURNALS